Untill now the javadoc comments were read by token/word. This works for most cases, but when comments are becomming complex because of the usage of asterikses (together with tabs and spaces) the result be slightly different then the input.
I've rewritten the lexer, where the reading is based on lines instead of tokens. This makes the result much more solid, but requires some special lexing.
One testcase had 2 javadoccomments in 1 line (no sourceformatting, just one line of code). To read untill the first javadocEnd I had to use regepx inside the lexer-tokenblock.
When this patch is validated we can probably solve several open issues which refer to the usage of javadocs.
|Resolution||Fixed [ 1 ]|
|Status||Open [ 1 ]||Closed [ 6 ]|
|Assignee||Robert Scholte [ rfscholte ]|